# Agent Harness

> Turn any domain folder of skills into a bounded agentic loop: a manifest builder inventories a domain's skills/tools/checks, a goal compiler turns a goal into a verifiable task plan (refusing vague goals with forcing questions), and a JSON-backed loop controller drives execute-verify-close with retry caps, controller-run verification (no verification theater), human escalation on exhausted budgets, and a close gate that refuses while any task is unverified or unwaived. Ships 3 stdlib Python to

## Facts
- Page: https://tashan.sh/capability/plugin-alirezarezvani-claude-skills-agent-harness
- tashan id: plugin:alirezarezvani/claude-skills/agent-harness
- Source: https://github.com/alirezarezvani/claude-skills
- Type: plugin
- Category: devtools
- tashan score: 34.0 / 100
- Adoption: 7.0
- Upkeep: not measured
- Freshness: 81.0
- Evidence coverage: 59% of the inputs this score can use
- Health: active
- Instruction depth: not yet graded
- License: MIT
- Official: no

## Install

```sh
/plugin marketplace add alirezarezvani/claude-skills
/plugin install agent-harness@claude-code-skills
```

## Security audit
Not scanned. We audit npm-published capabilities; this one has no npm package we can resolve, or has not reached the queue. This is not a clean bill of health.

---
Measured 2026-09-13 by tashan (https://tashan.sh) from public evidence. Scorer s5.